I wouldn't be sure of that: - if by efficiency you mean the speed of operation, then it's the network size that determines it - the input token / vector size

I wouldn't be sure of that: - if by efficiency you mean the speed of operation, then it's the network size that determines it - the input token / vector size is not dependent on the language